tamilocrx 0.1.0

Tamil OCR engine using PARSeq and CRAFT

Yes, good catch! ✅ You need to have a README.md file in your project root before building the package. It's used as the long description for your PyPI page (what people see when they visit your package).

---

✍️ How to Create README.md

In your project root (E:\tamilocrx), create a new file named README.md and add something like:

# tamilocrx

`tamilocrx` is a Python package that extracts Tamil text from images using deep learning-based OCR models like **CRAFT** for text detection and **PARSeq** for recognition.

## 🚀 Features

- Detect and extract Tamil and English text
- Based on PyTorch, CRAFT, and PARSeq
- Easy to use with Gradio demo

## 🔧 Installation

```bash
pip install tamilocrx

🧪 Usage

from tamilocrx.ocr import OCR

ocr = OCR(detect=True, enable_cuda=False)
result = ocr.predict("path/to/image.jpg")
print(result)

📦 Model Files

• parseq_tamil.pt – Tamil text recognition model
• craft_mlt_25k.pth – Text detection model